Unions and the Seniority “Threat”:

The impact of internal organization on unions' pension policy preferences

Karen M. Anderson Department of Public Administration Leiden University Box 9555 2300RB Leiden The Netherlands tel: +31 (0) 71 527 3893 fax: +31 (0) 71 527 3979 anderson@fsw.leidenuniv.nl

Julia F. Lynch CBRSS Harvard University 34 Kirkland St. Cambridge, MA 02138 USA tel: +1 617 495 5366 fax: +1 617 496 1636 jlynch@latte.harvard.edu

*** WORK IN PROGRESS: Comments especially welcome ***

Paper prepared for presentation at the 14th International Conference of Europeanists, March 11-13, 2004 at the Palmer House Hotel in Chicago.

We would like to thank Todor Enev and Henriette Litta for their tireless research assistance, as well as Bernard Ebbinghaus, Lucio Baccaro, and the staff of EIRO’s network of reporting institutions for generously sharing their expertise. We also thank Christa van Wijnbergen and Kent Weaver for their most valuable comments on an earlier draft.

Abstract

Throughout Europe, workforces are ageing, and the membership profile of unions is

becoming weighted towards older workers and pensioners. How do these changes affect the

positions that unions adopt with regard to welfare state reforms? This paper argues that how

pensioners are represented within unions, rather than how many pensioners there are in

unions, helps shape confederal positions on pension reform. Our theory suggests that peak-

level representatives of labor will be more likely to support pension reform proposals that

involve substantial modifications to the future benefits of current workers, as long as the

pensions of most current retirees are not threatened, if two conditions are met: (1) labor

movements are characterized by a high potential for pensioner influence at the confederal

level, and (2) confederal (peak-level) union structures exercise meaningful authority over

sectoral unions. Evidence from Italy and Germany confirms that the “graying” of union

membership per se is unlikely to be the source of union intransigence on pension reform.

One important policy implication is that strengthening the representation of pensioner

interests within unions may actually help, rather than hinder, the process of long-term

pension system reform.

1

I. Introduction

Throughout Europe, workforces are ageing, and the membership profile of unions is

becoming weighted towards older workers and pensions. How do these changes affect the

positions that unions adopt with regard to welfare state reforms? The dominant view,

espoused by many scholars and policy analysts, is that as union memberships age, a

“seniority bias” driven by the predominance of older workers and pensioners will prompt

unions to pursue political strategies aimed at preserving acquired rights and the pension

policy status quo (for scholarly work in this vein, see Anderson and Meyer 2003; Brugiavini,

Ebbinghaus, Freemen et al. 2001). But other research suggests that the impact of ageing

populations on welfare state policy may be more complicated. Pampel (1994) and Lynch

(2002, chap. 7) find that when neocorporatist decision-making structures are in place,

relatively aged populations actually result in an increased share of social spending dedicated

to young people. And Campbell and Lynch (2000) argue that pensioners’ organizations,

including pensioners’ unions, are often more reform-oriented in their claims on the pension

system than elderly voters are. This body of research suggests that the presence of large

numbers of pensioners in unions may, under certain circumstances, actually prompt

confederal unions to take more moderate positions on pension reform than they would

otherwise.

We enter this debate about the impact of ageing on social policy with the claim that

how pensioners are represented within unions, rather than how many pensioners there are in

unions, shapes confederal positions on pension reform. This paper focuses on two

dimensions of the organization of pensioner interests within union movements that seem

likely to affect unions’ pension policy preferences: whether pensioners are represented in the

sectoral unions of their former employment or in their own pensioners' unions; and the degree

of centralization of union federations. We argue that centralized unions with pensioners

2

organized in independent pensioners’ unions are most likely to support significant pension

system reforms.

Why do these organizational features internal to unions matter? Our argument makes

two central claims. First, different actors within unions want different things when it comes

to pensions. We focus on differences in preferences about pension reform between four key

kinds of actors: current workers versus current pensioners, and sectoral leaders versus peak-

level leaders. We argue that current pensioners and peak-level leaders are more likely to

support pension reform than either current workers or sectoral union leaders.

Second, the internal organization of unions plays an important role in determining

unions’ social policy preferences because different organizational structures privilege the

voices of different actors within unions differently. Union centralization (which implies

confederal authority over affiliate unions) clearly favors the voices of peak-level leaders. But

in all but the most determinedly neocorporatist systems, peak-level control over affiliates is

often incomplete, and fraught with uncertainty. We argue that strong independent pensioners'

unions that ensure pensioners a privileged voice at the confederal level can strengthen the

hand of pro-reform peak-level leaders and encourage them to take more moderate stances on

pension policy than they would otherwise be able to take. Of course, actors' preferences are

not independent of the institutions that channel them; we note in particular that the

institutional form of pensioner organization is likely to affect how pensioners perceive and

articulate their own interests, such that independent unions lead to not only a stronger voice,

but also a more pro-reform voice, for pensioners.

To illustrate our theoretical argument, we examine recent pension politics in Italy and

Germany, two countries with Bismarckian pension systems in which pensioners make up a

large share of union membership, but in which pensioners are organized quite differently

within the major labor unions. The case studies confirm that the organization of pensioners

3

within unions matters for the positions that unions take with regard to welfare state reform,

and matters in counter-intuitive ways. In Italy, tripartite concertation surrounding welfare

state reform strengthened the hand of the three main union confederations' confederal leaders,

who allied with powerful pensioners' unions to counterbalance anti-reform positions pushed

by sectoral unions. This alliance facilitated the emergence of a pension reform compromise

in 1996. In Germany, the severity of pension cuts included in the Red-Green government's

2000 reform proposal took unions by surprise. The Federation of German Trade Unions,

DGB (Deutscher Gewerkschaftsbund), tried to push for less drastic cuts in the midst of union

disagreement. Like the pensioners’ unions in Italy, the DGB moderated more oppositional

demands among its affiliates and facilitated a compromise. But because the confederal

leadership has lacked a strong pro-reform ally -- in the form of an independent pensioners’

union, for example, -- the demands of strong sectoral unions have tended to prevail, slowing

pension reform in Germany.

The paper is organized as follows: Section II begins with a brief literature review,

and then sets forth the theoretical basis for our claim that different forms of pensioner

representation within union movements affect the formation of union preferences on social

policy. Section III illustrates the argument with case studies of pension policy-making in

Italy and Germany. We conclude in Section IV by drawing out some implications of our

analysis for both the welfare state literature and the corporatism literature.

4

II. Theoretical overview: Institutions of pensioner representation and union

preferences on pension reform

A. An extremely brief literature review: “seniority bias” confronts institutionalism (and the

institutions win, naturally)

While most scholars would agree that unions have played an important role in the

development of welfare states, we still know remarkably little about the origins of unions’

social policy preferences, especially when retrenchment reaches the political agenda. As a

result, we have few conceptual tools for understanding the potential impact of ageing

workforces on union social policy preferences and behavior. The few studies that do

investigate the role of unions in processes of welfare state restructuring tell us that the

number of older union members is what matters most for union preference formation

(Brugiavini, Ebbinghaus, Freeman et al. 2001; Anderson and Meyer 2003; and, by

implication, Dickson and Shepsle 2001). The central argument of this line of analysis is that

the membership composition of unions plays an important role in shaping union social policy

preferences. When union membership becomes concentrated among retirees and older

workers, unions are likely to exhibit a "seniority bias" that leads them to prefer the

preservation of the pension policy status quo and to try to push the costs of reform onto

younger generations.

We think this is an incomplete approach to understanding union social policy

preferences for two reasons. First, we have reason to suspect that the elision of aging

workers and retirees into a single group from the standpoint of their interests in pension

reform is misleading. The seniority bias thesis assumes that current pensioners and workers

close to retirement share a common interest in preserving the pension system status quo. We

5

find this microfoundational assumption to be deeply problematic, and propose an alternative

set of assumptions in Section III below.

Second, the seniority bias approach assumes that the policy positions espoused by

union leaders are a simple function of the age structure of membership. Union leaders are

presumed to follow the preferences of the median member: as the median member ages,

union pension policy preferences tend to be oriented toward protecting the existing or

expected benefits of older union members. This approach overlooks the institutions that

structure interactions between sectoral unions and the union confederations to which they

belong, as well as between different union confederations.

There is ample evidence, though, that at the macro level, institutions matter for who

gets represented and what their interests are. Previous scholarship, not least in the "varieties

of capitalism" framework, emphasizes that macro-political and labor market institutions

matter for how the social partners come to perceive their own interests in the field of social

policy.1 The institutionalist approach rejects the assumption that the main players in the

political economy have social policy preferences that can be read unambiguously off of their

location in the productive structure. Instead, union and/or employer preferences and political

behavior are shaped by institutional factors such as policy legacies and the function of social

policies in the labor market. Our argument builds on this institutionalist insight about the

sources of union policy preferences. We shift, though, from a focus on the macro-

organization of the political economy to an investigation of how the micro-organization of

key actors contributes to their policy positions.

Contrary to the assumptions of the seniority bias thesis, we argue that there are

important differences in how unions are organized internally, institutional differences that

affect in important ways how the preferences of union members are aggregated. In other

1 See for example Anderson (2001), Manow (2001), Mares (2003), Swenson (2002).

6

words, aging union memberships do not translate automatically into peak-level opposition to

pension reform. In the following section, we lay out the logic underlying our claim that

institutions internal to unions – chief among them, the way that pensioners are organized and

the relative strength of horizontal (cross-sectoral) versus vertical (sectoral) structures – affect

union preferences concerning pension policy reform.

B. A theory of pensioner influence on union policy preferences

Pension reform entails losses for some groups and not for others. Our goal is to better

understand how peak-level union leaderships balance these losses and gains for diverse

groups within organizations as they develop policy preferences surrounding pension reform.

Under what circumstances will peak-level leaders support pension reforms that hurt some

members? Under what circumstances will they oppose reforms, even when they would help

some of their members? To answer these questions, we need to begin by understanding what

is at stake for various groups when pension reforms are on the table.

We focus here on public pension schemes in the Bismarckian mold, in which the

entitlement for pensions depends on employment history, the level of the pension is related to

the level of prior earnings, and in which both the replacement rate and the ratio of benefits to

employer and employee contributions typically varies with the sector of employment and/or

job classification. Beveridgean, citizenship-based pension systems, in which the basic public

pension is a universal, often flat-rate, entitlement financed via general revenues, clearly imply

a different set of concerns and priorities for both union leaders and union members. There

are no instances of independent pensioners' unions in countries with Beveridgean pension

systems, and in many of these countries few pensioners remain union members after

retirement (see Tables 1 and 2). The main advocacy organizations for pensioners in these

systems tend to be non-union groups, which exert pressure on tripartite actors in different

7

ways. For these reasons, we leave analyses of union decision-making in Beveridgean

systems for a future date. What, then, are the implications of Bismarckian pension systems

for the preferences of different actors within labor unions?

We hypothesize that different kinds of union actors -- pensioners, current workers

(prime-age workers in strong sectors, prime-age workers in weak sectors, and young

workers), sectoral union leaders, and confederal (peak-level) leaders -- have distinctive

preferences regarding pension reform. In general, we argue, pensioners and confederal

leaders are likely to be the most favorable towards pension system reforms proposed by

governments and employers, while current workers and sectoral union leaders are likely to be

less favorable. Within this latter group, important differences exist, with younger workers,

workers with less secure employment, and both workers and union leaders in sectors with

weak bargaining leverage more likely to favor some kinds of reforms. The institutions that

structure interactions among these groups are key to determining which group or groups'

preferences get expressed at the peak level, and also contribute to the formation of

preferences. Keeping in mind these basic outlines, we can examine in more detail the

interactions between preferences and institutions that concern us, and how these interactions

contribute to the propensity of confederal union leaders to support or oppose pension system

reform.

Let us begin by remembering why pension reform is on the table to begin with. The

major challenge confronting most public pension systems in Europe, be they Beveridgean or

Bismarckian, is some form of financial imbalance. Most public pension systems in Western

Europe use pay as you go (PAYG) financing, a defined benefit structure, and rely for much of

their financing on mandatory contributions made by employers and employees. In many

pension systems, the central government also provides a subsidy to the pension system, either

to compensate for revenue shortfalls or to finance the contributions of wage earners not

8

currently in dependent employment because of unemployment, sickness, childrearing, and/or

education. Financial imbalance occurs when the ratio of contributors (employed persons) to

beneficiaries (retirees) declines, due to increasing longevity, declining birth rates, high

unemployment, or workers exiting the labor force in large numbers through early retirement

or disability. It may also occur because pension rights have been granted politically without

providing for an inflow of resources necessary to fund them. Whatever the source of

financial imbalance, most governments are no longer willing or able to increase general

revenue subsidies to make up for declining payroll tax revenues as a share of total pension

spending. This means that restoring some semblance of financial balance usually requires tax

increases or benefit cuts, or a combination of both. In Bismarckian systems, the tax increases

are most likely to appear as increases in mandated employer and/or employee contributions.

To ease the pain of retrenchment, cuts and payroll contributions can be phased in for future

workers/retirees.

Bismarckian earnings-related pensions are based on a long-term promise by the state

(or social insurance administration) to pay deferred wages in the form of a retirement

pension. For all current workers, three things are important: the level of pension

contributions required to finance pensions, the level of future pensions, and the likelihood

that the pension system will deliver on its deferred wage promise at retirement. The level of

contributions affects take-home pay, but also the willingness of firms to hire new workers.

The level of future pensions determines the likely reward for employment, but also the

sustainability of the pension system. The likelihood that the pension system will deliver on

its promise of deferred wages, clearly, depends on the financial and political stability of the

system as a whole. Different groups of union members prioritize these goods differently,

depending on the distance from retirement, the anticipated level of their own pensions, and

other considerations. Let us examine some of these differences in perspective in more detail.

9

Preferences for pension reform

Current pensioners. We can begin with a group that is in fact not, by and large,

concerned with the level of pension contributions, the level of future pensions, or the

likelihood that the pension system will deliver on its promises for current workers. This

group is made up of current pensioners. We argue that unlike current workers, and perhaps

counter-intuitively, current pensioners have little to fear, and sometimes much to gain, from

pension reform. They have little to fear because their sheer numbers and electoral clout

ensure that even if unions support pension reform at the peak level, current benefits are

unlikely to be cut. Notions of personal contributions and/or deferred wages inherent to

earnings-related pensions also ensure that current pensioners are likely to be seen as having

"earned" their pensions, making cuts to current pensions politically difficult. Some cuts to

current benefits have occurred in recent pension reforms in Continental Europe, but we argue

that these kinds of cuts are likely to be the exception to the rule.2

Pensioners may in fact welcome pension reform, as long as the reforms involve

cutting future benefits or raising current contribution rates, and not cutting current benefits.

Current pensioners, even union members, are unlikely to care much about the level of

benefits for future pensioners beyond a vague feeling of solidarity with future generations of

workers. On average, current pensioners enjoy benefits lower than the benefits that the

average current worker on the eve of retirement can expect. Not only have pensions gotten

more generous with the accumulation of union victories over the post-war period, but the

greater longevity of women relative to men means that a higher proportion of current

pensioners than of current workers are women, many of whom have weaker pension

2 For example, Swedish pension reforms in the early 1990s included temporary cuts in current pensions, as have reforms in the Netherlands and Germany in 2004. The essential point here is that cuts in current pensioners are rarely permanent; they are typically achieved by temporary suspension of indexing, as was the case in the examples mentioned above.

10

entitlements than men. So current pensioners may not see the need to defend at all costs

benefits for future retirees that are more generous than those they receive themselves.

But while pensioners may not care much about the level of future benefits, they care

deeply about the sustainability of the public pension system. Recent retirees can look

forward to an average of about twenty years as pensioners, and many will have no other

sources of income during that time.3 Concerns about sustainability may lead current

pensioners to support cuts to future benefits or increases to current contributions in order to

ensure financial balance in the pension system. Current pensioners may also be willing to

support a reduction (or at least non-expansion) in the size of the pension system because they

depend more than do current workers on a wide range of public benefits, such as health care

or social services, which may be crowded out by pension spending. For all of these reasons –

they need not fear cuts to their own benefits, they care little for cuts to future benefits, and

they welcome some collateral benefits of restraining pension spending -- , we expect

pensioners to be more favorably disposed towards pension system reform than most groups

of current workers. More formally, we can express their preference ordering as follows:

cut future benefits = increase contributions = SQ > increase state subsidy .

(Cutting current benefits does not appear in the preference ranking here or in those for other

groups below because, as we argue above, it is unlikely to be a politically viable option.)

Young workers, paradoxically, share many pensions reform preferences with the

elderly; they certainly care less about future benefit entitlements than their prime-age

colleagues do. The preferences of young union members are of course affected by the human

propensity to discount benefits and losses that accrue far in the future. But they also know

that pension system reform of some kind is inevitable, and that the politics of pension reform

make it likely that the young will bear the brunt of any phased-in solution. Young people 3 This is especially true because the earnings-related public pension systems common in continental Europe tend to both encourage early exit from the workforce, and discourage the development of private pension alternatives.

11

vote at lower rates than older people do, and high levels of youth unemployment in

continental Europe mean that young people are less likely to be represented in unions. This

political weakness of young adults, combined with the compelling rationale of phasing in cuts

for those groups who have the longest time in which to adjust their own behavior to

compensate, mean that pension reforms are almost guaranteed to hit the youngest workers

hardest.

Increased state subsidies to shore up the pensions system may thus seem desirable to

young workers, but only if they ignore the prospect of having to pay down the public debt

later. For the same reasons, younger workers are likely to resist higher contributions now in

order to pay for benefits that they are unlikely to receive. Further, they may see advantages

in pension system reform if it contributes to increasing the number of permanent jobs by

reducing non-wage labor costs. This will be especially true in countries where unionization

rates among workers on short-term contracts are high. However, even young people with

"permanent" jobs may be more interested in labor market performance than in preserving

pension benefits, either because they do not wish to be locked into their current jobs, or

because of their personal contact with school-mates and siblings who have not been able to

find employment. The preference ordering of young workers can therefore be represented as

cut future benefits > or = increase state subsidy = SQ > raise contributions .

Current prime-age workers in strong sectors (who can afford to pay higher

contributions) with generous future pensions are likely to favor the pension status quo

because of their interest in claiming future benefits at least as generous as current benefits.

These workers are willing to pay high contributions in order to receive generous future

benefits. They are likely to oppose pension cuts, preferring instead to pressure the state to

provide more funds for the pension system or to broaden the revenue base. This group is

likely to accept increased contributions if the first two alternatives fail. This group is not

12

likely to favor (current or future) pension benefit cuts because this will affect their own future

pensions. Strong sectoral unions usually represent this type of wage earner and they are often

the most vocal opponents of reforms, since they have managed to secure the most generous

pensions and they do not want to see these pensions reduced. This group's preference

ordering looks like this:

SQ > increase state subsidy > raise contributions > cut future benefits .

Workers in weak sectors. Not all workers, even unionized workers, have well-paid,

stable jobs that provide generous pensions. Workers in weak sectors are less able than more

privileged fellow union members to accept increased contributions, and they may be less

willing to do so in the name of safeguarding a pension system that benefits others

disproportionately. By definition, earnings-related pension schemes reproduce income

differentials in retirement, and the occupational fragmentation of pension schemes often

magnifies these differences. Lower income workers or those in unstable unemployment may

thus favor pension reforms that reduce future benefits if they foresee cuts to higher-end

benefits rather than to the more modest pensions that they receive. The "vice into virtue"

(Levy 1999) flavor of many current reform agendas in fact targets the pensions of privileged

groups in order to restore financial balance to pension systems or to improve the benefits of

less privileged groups. Under such circumstances, workers in weak sectors may in fact

support pension reforms that cut future benefits. This group's preference ordering are a bit

harder to determine in the abstract, since much will depend on whether reforms attack only

high-end pensions or reduce benefits across the board. Assuming a “vice into virtue”

framework in which reforms target higher pensions, the preference ordering for workers in

weak sectors will look like this:

increase state subsidy > cut future benefits > SQ > increase contributions .

13

If, however, the reform process becomes dominated by strong sectoral unions and cuts are not

concentrated among the best-off workers, preferences among workers in weak sectors could look

more like this:

increase state subsidy > SQ > increase contributions > or = cut future benefits.

Union leaders. It is useful also to consider the likely preferences of sectoral union

leaders and peak-level union leaders with regard to pension reform. Sectoral union leaders

are likely to take defensive positions on pension reforms that affect their members directly,

since the legitimacy of a sectoral union leader depends first and foremost on his or her ability

to protect the particular interests of workers in that sector. Sectoral leaders are thus expected

to have interests congruent with those of the majority of the members of the unions they

represent. Leaders of sectoral unions in strong sectors with ageing workforces will likely be

intransigent on pension reforms that threaten benefits for workers who are approaching

retirement, while leaders of sectoral unions whose members have weaker predicted pension

entitlements are more likely to accept reform proposals that seek to redistribute benefits while

cutting aggregate costs.

The preferences of confederal leaders are difficult to specify in the abstract, since

institutions that connect them to their constituencies in different ways may result in different

priorities. In general, though, we can say that confederal leaders must reconcile two

conflicting sources of pressure that may threaten their own position as leaders. On the one

hand, and to secure their position internally, union leaders must work to protect the expressed

interests of workers in the strongest sectors. Particularly in confederations where peak-level

authority is weak, confederal leaders may adopt anti-reform positions in order to maintain the

support of strong sectors. On the other hand, adopting pro-reform positions often helps to

secure peak-level leaders' roles as privileged interlocutors with governments and employers.

And because confederal leaders have a mandate to represent the interests of all of their

14

members4, not just those with the most to lose from pension reform, they may perceive a

responsibility to sublimate the disparate interests emanating from sectoral unions into a

reasonable package of pension reforms in order to ensure some greater good (eg. long-term

sustainability, lower taxes, more equitable distribution of retirement income). So confederal

leaders may have a "natural" inclination to adopt conciliatory positions on pension reform,

especially if confederal authority over affiliate unions is strong enough to preempt threats

from below.

So far, the discussion has centered on the “natural,” or “unorganized” preferences of

groups of workers and pensioners – the union base – and different levels of union leadership.

We turn next to investigate how different modes of union organization affect the preferences

of some union members (particularly retired members) and union leaders, and privilege the

voices of some groups over others.

The Organization of Pensioners.

There are four main ways in which pensioners may be organized within unions:

1. Pensioners leave their sectoral union and become members of distinct pensioners’ unions that have status within the confederation equivalent to other sectoral unions.

2. Pensioners remain organized within the sectoral union of former employment, but have some form of independent representation at the confederal level (similar to many youth, women’s groups)

3. Pensioners remain organized within the sectoral union of former employment, but have no independent voice at the peak level

4. Pensioners are not permitted to remain union members.5

Each of these organizational configurations shapes both the expressed preferences of

pensioners, and the likelihood that these preferences will be adopted by peak-level leaders.

Let us examine the implications of pensioner organization more closely.

4 This mandate sometimes extends to a perception of peak-level responsibility for the welfare of all workers, unionized or not; or even the working classes as a whole. 5 See Künemund, Neckel and Wolf (1994) for a similar typology.

15

Pensioners’ unions. The requirement that union members relinquish membership in

their sectoral union with the option of joining a pensioners' union has important implications

for union politics and the aggregation of pension policy preferences. The first and most

obvious implication is that the change in union membership strengthens the union member's

identity as a pensioner rather than as a former wage earner in a specific sector (Künemund,

Neckel and Wolf 1994). If former members of different category unions are combined into a

single pensioners' union, sector-specific issues should decline in importance compared to

interests shared by all members of the pensioners' union. Members of pensioners' unions

should care less about issues related to their former sector of employment (wage levels,

working conditions, etc) and more about retirement-related issues such as the level of pension

benefits, pension indexing, or access to social services.

The creation of a new, more “encompassing” (see Olson 1982) identity as pensioners

can also facilitate a recognition by pensioners who are union members that maximalist claims

on the pension system may not help current pensioners in the aggregate. Current pensioners

are a mix of relatively well-off recent retirees with pension rights accumulated over a full

career; retirees with pensions that are less generous because of lower wages, less stable

employment, and/or a weaker bargaining position of their sector of former employment; and

still poorer, older, often female retirees whose pensions are at a minimum level. Union

positions that defend the privileges of recent retirees in strong sectors while forcing cuts in

the provision of services to other, less well-off retirees are less likely to be endorsed by

members of independent pensioners’ unions who share a common identity as pensioners.

A second important implication is that pensioners' unions, like all complex

organizations, are likely to develop a professionalized leadership cadre with expertise on

issues important to the union and with a perception of their own stake in the sustainability of

the pension system. This means that the pensioners' union officials may be willing to support

16

reforms that protect the long-term solvency of the pension system even if it means advocating

somewhat lower benefits in the short-to-medium term (see Campbell and Lynch 2000).

Contrary to conventional wisdom that elderly interest groups are uniformly powerful

opponents of pension reform, then, we expect that pensioners’ unions may actually be forces

in favor of reform, especially if benefits for current pensioners are protected.

Pensioners integrated into existing category unions. The most common form of union

membership for retirees in Continental Europe is as members of the sectoral union

representing the former sector of employment. This type of union membership is usually

offered to retirees for a reduced union fee and often includes inducements such as the

opportunity to buy inexpensive group insurance. A key dimension of this type of

organization is whether or not retired members have voting rights and their own

representation at the union level and confederal level. Unions may have many retired

members but few or no channels for pensioners to take part in and influence activities. This

means that retirees remain union members largely to maintain contacts with former

colleagues, to benefit from union-provided services or to take part in recreational activities

provided by unions. Here the institutionalization of pensioner influence within the union is

weak. At the other extreme, there may be institutionalized channels for pensioner influence

within unions, such as the right to form a separate pensioners group with voting rights and

representation at the union and/or confederal level. In this context, retired union members are

organized within the union and have regular opportunities to discuss and formulate positions

related to their status as retirees, including pension issues. Here, the institutionalization of

pensioner influence is strong.

The implications of retaining pensioners as members of sectoral unions, rather than as

a separate category union, are threefold. First, when retirees remain members of sectoral

unions, they are likely to retain their identity as a member of that sector of employment, and

17

less likely to develop a separate union identity as a pensioner. This tendency should be

stronger when retired union members have few or no voting rights and do not have separate

representation at the confederal level. Second, when pensioners and workers are combined

in sectoral unions, workers’ interests will prevail over those of pensioners both within the

sectoral union and at the peak level. There are usually more workers than pensioners in any

given union, and workers are closer to core mission of unions. To this end, mechanisms to

reduce pensioners’ influence are typically put in place when pensioners form a majority of a

union’s membership. Third, the absence of independent organizations of and for pensioners’

makes it more difficult for a professionalized pensioners’ leadership cadre to emerge. In this

case, the confederal union leadership is likely to remain the sole voice for the solvency of the

welfare state as a whole, and it may be difficult to maintain such a policy position when

strong sectoral unions reject it.

The previous discussion suggests the following propositions. First, when pensioners

have no independent voice at the confederal level, this is likely, ceteris paribus, to lead to a

predominance of category-specific interests when pension reform is on the agenda. Second,

if pensioners have their own unions with representation at the confederal level, they can push

their own (often more moderate) positions, and the confederal leadership will be more likely

to adopt a pro-reform agenda as long as current benefits are protected. Finally, if pensioners

do NOT have their own unions, but do have an independent voice at the confederal level, we

would expect a “medium” result, in which the confederal leadership adopts some pro-reform

policy positions, but meets strong resistance from category unions.

Centralization.

The centralization of confederal authority also shape the way that union leaders

respond to pressure from unionized pensioners. Concertation may be seen as a special

18

instance of centralization in which the importance of the peak level vis-a-vis affiliates is

temporarily increased due to the necessity for unions to speak with a single voice to the other

social partners and to government. Both concertation and confederal authority over affiliate

unions magnify the effects of pensioner organization identified above, by increasing the

importance of encompassing, cross-sectoral union structures like the confederal leadership

and pensioners’ unions.

When the peak-level leadership of a union wishes to make a deal with the government

and/or employers for pension reform, it needs the support of affiliates. If an independent

voice for pensioners exists, and if that voice favors reform, confederal leaders will listen --

and the support provided by the pensioners' unions may enable confederal leaders to defeat

reform opponents in other sectoral unions. During periods when there is no concertation, or

if power within the confederation is weakly centralized, sectoral unions are likely to prevail

over the confederal level. Under those conditions, independent pensioners’ unions will not

have much impact, since their voice will have no privileged ear at the confederal level, and is

likely to be drowned out by the voices of more militant sectoral unions.

To sum up the argument so far, we have argued that the amount and kind of pressure

that a graying union movement puts on confederal leaderships with regard to pension reform

depends on how pensioners are organized within unions. We have argued further that the

broader environment within which unions act will also affect the way that the leadership

responds to pressure from pensioners within their ranks. During episodes of tripartite

concertation peak-level bodies within unions are empowered and may pursue pension

policies that are not in the short-term interests of the strongest category unions.

Uncoordinated bargaining, on the other hand, encourages individual category unions to

defend particularistic interests in the pension system status quo, in a dynamic similar to

‘leapfrogging’ in the area of wage bargaining.

19

Our theory suggests that labor movements characterized by (1) a high potential for

pensioner influence on union demands and (2) strong confederal authority over category

unions will predispose confederal leaders to support pension reform proposals that involve

substantial modifications to future benefits, as long as the pensions of most current retirees

are not threatened. The potential for pensioner influence, the first of our two key dimensions,

is reflected in three aspects of union organization:

1. The presence or absence of independent pensioners’ unions 2. The presence or absence of a pensioners’ current with official voice. 3. The numerical weight of pensioners within the confederation (a critical mass above a

certain threshold – say 5 percent? – seems necessary to qualify a country for consideration here)

We emphasize potential for influence, since we expect that the degree of confederal authority

is a limiting factor on the actual influence of pensioners. Without effective representation,

pensioners have no "voice." But without an effective confederal leadership, independent

pensioners' unions have no "ear." Assuming a critical mass of pensioners exists within the

unions, independent pensioners’ unions give the highest potential level of influence, an

organized current with confederal representation the next-highest level, and no organized

representation the lowest level of potential influence for pensioners qua pensioners.

The numerical weight of pensioners in union confederations varies dramatically, with

figures in the low single digits for Denmark, the UK and Ireland, and Portugal, and figures

above 50 per cent for some confederations in Italy and France. The figures reported in Table

1 below represent our best estimates.6 Table 2 summarizes what we know about the

organization of pensioners within European labor union confederations. (Many lacunae 6 A number of factors could account for variations in the numerical and proportional weight of pensioners within unions across the European continent. Population ageing is farther along, and proceeding more quickly, in some places than in others. So too is the ageing of the workforce per se, since labor markets very in their openness to young entrants. Union efforts to recruit younger members vary across countries, as does the success of these efforts. Finally, unions in different national contexts treat retirees differently for the purposes of membership, dues-paying and the like. Accurate figures on the number of pensioners in unions are difficult to obtain, not least because in an era of declining union membership, many confederal and category unions are hesitant to reveal how many of their members are in fact not active workers. At the same time, pensioners’ unions have an incentive to inflate their membership figures in order to demonstrate their power and relevance. So the direction of the likely error depends on the source of information, assuming that one can find any information at all.

20

remain; feedback from knowledgeable parties is most welcome.) Independent pensioners’

unions exist within the major union confederations in France, Italy, Portugal, and Spain.

Among the countries in which pensioners remain members of the category unions in their

former sector of employment, Austrian, Dutch, Luxembourgish, and British (?) unions give

pensioners formal representation and sometimes even voting rights in confederal-level

decision-making bodies. Elsewhere (e.g. Germany, Belgium?), pensioners have no

independent voice at the confederal level. Clearly formal representation at the confederal

level is no guarantee of influence over confederal policy positions, but it may indicate a

higher level of pensioner influence over peak-level policy positions. The degree of

condederal authority over affiliate unions in reported in Table 3. The degree of centralization

can change quite dramatically over time, as unions become engaged and then disengaged

from peak-level tripartite concertation over social policy. Visser's (1990) summary measure,

which we report in Table 3, should be taken as something of an average over time, and not as

an indicator of the level of confederal authority at any given point in time.

Summarizing the basic characteristics presented here, national patterns of union

organization in countries with Bismarckian pension systems can be arrayed on two main

axes: the potential influence of pensioners at the confederal level (combining numerical

weight and the organization of pensioners), and the centralization of authority in confederal-

level union structures.

21

Union centralization Potential for pensioner influence

Low High

Low (sectoral unions, few pensioners)

AustriaBelgium

Germany Portugal

High (own unions, many pensioners)

Spain (most of the time) France Italy 1975-1990

Spain (Toledo)

Italy 1990-2000

All other things being equal, which they most assuredly are not, we would expect in countries

in the lower-right quadrant to see confederal union leaders adopt positions favorable to

pension system reforms that (a) do not cut current benefits and (b) increase pension system

equity while cutting costs among those countries in the lower-right quadrant. Note that one

implication of this framework is that increasing union centralization (perhaps via a strong

commitment to tripartite resolution of the pension system crisis), rather than decreasing the

influence of pensioners, may be the key to spurring reforms in systems like France in the

lower-left quadrant. Conversely, giving pensioners strengthened organization within unions

may help bolster the reformist impulses of the moderately- and highly-centralized

confederations in the countries populating the upper half of the table.

A rigorous test of these claims clearly requires accounting for other potentially

important aspects of interest organization, the pension system, and the politics of pension

reform (union density, the presence of alternative representatives of pensioners’ interests such

as political parties or non-union advocacy groups for the elderly, the organization of the

pension system, the financial sustainability of current pension provisions, the array of reform

alternatives on the agenda, the presence of other issues that could serve as wedges or

tradeoffs, etc.) For the moment, we attempt merely to control for some aspects of the union

22

and broader political economy environment with the selection of two illustrative case studies,

Italy and Germany.

III. Testing the Theory. Italy and Germany.

Germany and Italy are good case studies to test our central theoretical arguments for

two reasons. First, the similarity of the pension systems in the two countries (public, PAYG,

earnings-related benefits) and the high number of pensioners relative to current workers in

unions in both countries allows us to control for the influence of these two variables. Second,

there are significant differences in the structure of pensioner organization in unions and the

degree of confederal authority over affiliates in each country.7

Recent rounds of pension reforms in both Italy and Germany have been well-

documented, so the following case studies will reveal few new details, but rather focus on

explaining where we think union preferences came from. Why and how were pensioners

and/or current workers able to get their demands to be the union's demands?

A. Italy

Italian labor unions are organized into three main confederations, the CGIL, CISL,

and UIL. In each of these confederations, a national-level confederal leadership coordinates

the activities of category and local-level unions. Each of the three major union

confederations in Italy has a pensioners’ union that occupies a position within the union

confederations analogous to category unions like metal workers or public sector employees. 8

The three major pensioners’ unions, which were established just after World War II,

have grown in tandem with an enormous and highly complex pension system. The Italian

pension system is primarily a public pay-as-you-go system fragmented along occupational 7 Note that for even an initial “proof” that organization of pensioners matters, we’d want a third case that holds confederal authority and everything else constant, and lets pensioner organization vary. 8 Many of the smaller so-called “autonomous unions” are organized in the same way.

23

lines. The system is funded by a combination of contributions from employers and

employees, and ad hoc state contributions. Benefit levels vary by sector due to different

contribution rates, indexation policies, retirement ages, etc. Before the reform of 1992, the

Italian public pension system boasted 47 separate funds for different segments of the

population (Baccaro 1999, 128). These funds are managed by public agencies governed

jointly by representatives of the state, employers, and labor unions.

The complexity and differentiation of the Italian public pension system has resulted in

a generalized perception that the system is inefficient, costly, and unfair. At the same time,

while virtually all observers agree that the system is in desperate need of a cure, analysts

disagree about the precise cause of the pension system’s maladies. One view identifies

electoral clientelism as a key to explaining the proliferation of expensive pension provisions

for micro-segments of the population (see for example , Ferrera 1984, Lynch 2001, Maestri

1994). Other analysts, however (Baldissera 1996b, Regonini 1990 and 1996), argue that it is

the social partners whose demands and concessions in the area of social policy are said to

have determined the shape of the Italian pension system, since pension legislation is the

product of negotiations among iron triangles of policy-making involving organized labor,

organized capital, and government agencies and ministries.

This second view, which implies that unions bear prime responsibility for the

pathologies of the pension system, has gained currency in recent years. In part this shift in

public and expert opinion is due to the dramatic collapse in 1992 of the parties most heavily

implicated in clientelist practices: if the parties no longer exist, how can they be blamed for

the pension system crisis? But there is also more than a grain of truth in the contention that

Italy’s unions have defended and even sought to expand expensive pension provisions.

Individual category unions and at times the confederal leadership have indeed defended many

of the “earned rights” of various constituencies, and played a major role in organizing the

24

national strike and public demonstrations that scuttled Prime Minister Berlusconi’s pension

reform plan in 1994, as well as more recent (and more half-hearted) reform proposals.

Labor unions in general, and the pensioners’ unions in particular, have borne the brunt of

public criticism for a welfare system increasingly characterized as benefiting pensioners

unfairly at the expense of the young.9 Pensioners’ union officials complain of public

discourse that “demonizes” the pensioners’ unions and “accuses them of an ‘assistentialism’

damaging to younger generations”10 (Acoccella 1999, 58).

The increasing blame placed on pensioners’ unions for Italy’s pension system woes is

partly a result of the increased visibility of pensioners’ unions. By 1993, the number of

pensioners surpassed active workers in the CGIL, the largest of the three main

confederations. In 1999 pensioners made up 55 per cent of the CGIL, 50 per cent of CISL,

and 25 per cent of UIL, for a total of 48 per cent of big three membership (Pedersini 2000)11.

Membership dues contributed by the pensioners’ unions to the confederal coffers made up a

substantial portion of the peak-level budget, giving pensioners’ unions claim to not only

political and organizational but also financial resources of the confederations (Regonini 1990,

352). The SPI-CGIL, for example, owns the very building that houses the CGIL

headquarters in Rome.

Italian pensioners’ unions are large, and they control valuable resources of the

confederations they belong to. But that does not necessarily translate into pressure for

confederal unions to defend the pension system status quo. Pensioners’ unions in Italy do

9 See for example opinion pieces in major news outlets by Manghi (1994), Ferrera (1996), Massarenti (1997), Rossi (1997), Cazzola (1998). 10 In the Italian debate “assistenzialismo” refers to unearned privileges, as opposed to earned rights. 11 These very large membership figures are due in large part to the Italian system of patronati – non-profit organizations, often linked to political parties and unions, that assist citizens in making requests to the public administration for pensions and other social benefits. The patronati are reimbursed by the state for each request that they bring. INCA, the patronato associated with CISL, made 8 million such requests in 1994 (Regonini 1990, 82). The patronato system enhances pension union membership because when a person makes a request through a union-sponsored patronato, he or she is often automatically enrolled in the pensioners’ federation of the union. One former union official I spoke with told me that his mother had inadvertently become a member (and was paying dues, however minimal) of all three major pensioners’ unions, having gone to all three patronati for help at various times.

25

have real influence at the confederal level, aside from their financial contributions, as we

describe below. This influence is even further enhanced during periods of tripartite

concertation or when the confederal leadership for other reasons is particularly strong vis-à-

vis category and territorial union structures. But the pensioners’ unions’ demands are not

oriented towards defending the current pension regime at all costs. Pensioners’ unions in

Italy have generally supported the confederal leadership in its pursuit of pension system

reform, as against the particularistic claims of some unions representing current workers.

Organization of pensioners within the unions.

The pensioners’ unions, like other category unions, have representation and voting

rights at the confederal level in Italy. But the large numbers of pensioners raises troubling

issues of internal representation for the confederations, which conceive of themselves, after

all, as representatives of Italy’s workers. For example, in 1999, SPI-CGIL, which controlled

55 per cent of the confederation’s membership, was the only affiliate that organized more

than 10 per cent. Even the biggest, strongest category unions could not come close to

matching the numerical weight of the pensioners’ union.

Along with large numbers (and the financial and organizational resources that go

along with them), the pensioners’ unions have an additional source of potential strength

within the confederations. Unlike active workers, whose demands may differ by

occupational sector or job status and run the gamut from working conditions to wage levels to

training programs, pensioners’ demands are relatively uncomplicated: they want decent

pensions and social services. According to Pedersini (2000), “if no ‘corrective’ mechanisms

are put in place, these interests could easily prevail in drawing up the union confederations'

strategies, all the more so given the importance of pensioners' unions in total membership.”

26

Two such “corrective mechanisms” are in place. First, the major confederations have

instituted rules that give pensioners’ unions less formal representation than they would have

by virtue of their size alone. Very large unions (i.e. pensioners’ unions) are limited to 25 per

cent of the voting delegates in confederal bodies like the national congress and the confederal

leadership committee (Pedersini 2000). Second, the pensioners’ unions have been

encouraged to become the confederations’ in-house specialists on issues related to the welfare

state, in order to “control potential overlap and conflict with sectoral unions' demands”

(Pedersini 2000). Article 12 of the CGIL’s constitution, for example, guarantees the SPI-

CGIL the right to propose language when the confederation takes a position of welfare state

issues. The SPI also “participates in the CCIL delegations which negotiate on the pensions

system, welfare services and the health system” (Pedersini 2000).

But such strategies to insulate the union confederation from pressure from large and

powerful pensioners’ unions have not been wholly successful. The pensioners’ federations

have become indispensable sources of expertise in the field of social policy, in part through

their cooperation across confederal lines. The early 1970s marked the beginning of close

cooperation among the three largest pensioners’ unions, and the inauguration of the resulting

unitary federation of pensioners’ unions’ as an independent interlocutor with the government.

The April 1975 government-union agreement indexing pensions to the growth in contracted

wages was signed by the Secretary General of the Federation of the three largest pensioners’

unions, rather than by the leaders of the confederal unions, as had always been the case

previously (FNP 1992, 11).

As the pensioners’ unions gained experience and independence, they gradually

became a leading source of social policy expertise for the union confederations. But rather

than using their new role to push for policies in support of workers’ acquired rights in the

pension system status quo, the pensioners’ unions began increasingly to adopt a role as

27

defenders of the welfare state more broadly. The strategy of giving authority over welfare

state issues to the pensioners’ unions, which Pedersini (2000) interprets as a strategy for

keeping the pensioners from dominating in other areas, has resulted in a leadership that is

assertive and confident in its claims surrounding the pension system.

These claims, however, do not typically mirror the claims of the largest and most

powerful category unions representing active workers. Campbell and Lynch (2000) argue

that the image of “voracious” pensioners’ unions defending the rights of the elderly at the

expense of the rest of society is misleading. Pensioners’ unions are in some regards more

moderate on the subject of pension reform than are category unions representing current

workers.

Pensioners’ unions were indeed formed specifically to advance the interests of former

workers and address the problem of poverty among elderly pensioners, and through the mid-

1970s a key demand of the unions was for bigger, better pensions for the elderly. By the

mid-1970s the basic goals of linking pensions to the cost of living and providing a non-

contributory “social pension” for elderly people with no other means of support had been

achieved. But pensioners’ unions weren’t all-powerful. Improved “minimum” pension

benefits12 and a new universal pension system, other goals of the pensioners’ unions,

remained elusive. And the pensioners’ unions were involved neither in the establishment of

seniority pensions nor in the setting of the low statutory retirement age, two features that,

more than any others, account for the very high cost of the Italian pension system (Lynch

2002, chap. 6).

The pensioners’ unions have a history of fighting for the losing side in Italian pension

reforms. They supported the unsuccessful February 1968 accord that would have raised the

retirement age for women, eliminated the possibility for people to receive pensions if they

12 Pension benefits for workers with some contributory history are topped up to a certain level in the absence of a minimum contributory history of 15 years. This level is the “minimum” pension.

28

were still working, and raised contribution rates, all in return for guaranteeing coverage to

some segments of the uninsured and linking the level of pensions to final earnings. This

accord failed to become law after opposition from industrial workers’ unions. As the unified

federation of the three leading pensioners’ unions asserted its claim to speak for confederal

unions on social policy issues in the mid-1970s, the positions that it staked out continued to

diverge from those adopted by the category unions. For example, in 1976, well before other

unions (or political parties for that matter) had taken up the call for pension reform, the FNP-

CISL warned of eventual collapse unless the pension system was reformed substantially

(FNP 1992, 12). The confederal leadership, supported by pensioners, backed the 1978 Scotti

accord, which was again scuttled by opposition from sectoral unions who would have been

hurt by it.

More recent policy proposals by the pensioners’ unions, in the midst of major societal

debates over pension reform, have demonstrated a similar willingness to advocate pension

reform. The yearly unified policy documents put forth by the unitary federation of

pensioners’ unions in the 1990s express consistent support for proposals that would limit

“seniority pensions,”13 early retirement pensions, and raises for the highest public old-age

pensions. Despite the pensioners’ unions’ active role in constructing the pension system in

the late 1960s through mid-1970s, pensioners’ unions supported the failed pension reforms of

1978-9 (Scotti), as well as the successful ones of 1992 (Amato), 1995 (Dini), and 1996

(Prodi).14

13 Pensioni di anzianita’ allow workers to retire at full pension after a certain number of years of contributions even if they have not reached the retirement age. 14 The 1994 reform attempt by Berlusconi, which threatened to reduce benefit levels for current pensioners by adjusting the indexation mechanism, did not receive the support of pensioners’ unions. Even so, pensioners’ unions were not out of line with the stance of the confederal leadership, which despite being marginalized by the Berlusconi government during the reform process, “declared repeatedly that they were not necessarily opposed to reform” (Baccaro 2000, 9).

29

Why pensioners’ unions are moderate on reform.

There are a number of reasons why we should not be surprised that Italian pensioners’

unions do not generally act as staunch defenders of the pension system status quo, echoing

the logic we laid out in Section 3 of this paper. In the first place, union representatives of

Italian pensioners may support pension reform because their constituents support it. For

example, a 1996 national survey revealed that workers in their 30s and 40s were far more

defensive of the pension system status quo than were pensioners (Baldissera 1996a). A

referendum called by the confederations to solicit workers’ and retirees’ opinions of the

proposed 1995 reform proposal garnered support from 91 per cent of pensioners. Only 58 per

cent of active workers voted in favor of the reform, and in sectors (e.g. metal working,

transport, teachers and university professors) where workers enjoyed extensive pension rights

that would have been most affected by the reform, a majority opposed the reform proposal

(Baccaro 1999, 150).

Current pensioners have less to fear from reform proposals than do current workers,

since their sheer numbers, and hence electoral threat, discourage politicians from advocating

reforms that would impinge upon the privileges of current beneficiaries. In addition, the

Italian Constitutional Court has consistently and strongly defended the notion of “acquired

rights” like pension benefits, overturning laws that deny promised benefits even in the name

of reform, e.g. of the disability pension system. As one Italian pension system reform expert

put it, “you need to worry about retrenchment if you're still active and about to retire, but if

you pass the mark you can relax” (Baccaro personal communication).

As a result, most Italian pension reform proposals in recent years have avoided

seriously impinging on current benefits in favor of cuts to future pensions15. And, not

surprisingly, the pensioners’ unions have found it easier to support reforms that spared

15 The Prodi reform did however drop indexation for the very highest pensions.

30

current pensioners while pushing cuts onto future generations of retirees. Meanwhile, the

union categories and professional associations that most benefit from generous seniority

pensions have lobbied actively against reforms (Regonini 1987).

Pensioners’ unions may also be relatively sanguine about pension reform because

current pensioners on aggregate have more to gain than current workers do from shifting

welfare state resources towards non-pension expenditures. Pension rights have gotten better

for most Italians in the last decades, and so the current generation of pensioners suffers most

from the very low entitlements granted to many women and other precarious workers in

previous decades. We have been unable to come up with useful descriptives of the average

pension rights of current pensioners versus workers within say ten years from retirement. But

an educated guess would put the numbers of current pensioners receiving relatively low

widows’, “minimum,” and “social” pensions well above the proportion of such low-end

pensions in the unionized population near to retirement. We do know that women make up

more than half of the membership of the SPI-CGIL (SPI-CGIL 1996), while they make up

much less than half of the unionized workforce in Italy. So even using this rough

approximation, we can gather that pensioners’ unions are likely to represent the less-

privileged elderly as well as richer pensioners, whereas sectoral unions in Italy have

increasingly come to represent primarily the highly protected core workforce close to

retirement age. It is not surprising, then, that pensioners’ unions are less protective of

“acquired rights” for the best-off pensioners, and more willing to accept pension system

reforms that offer increased social services or health benefits for the elderly in return.

Pensioners’ unions have also developed a professionalized staff with a stake in the

preservation of the welfare state as a whole. Pensioners’ union mass-audience publications

and publicity materials reveal an almost obsessive desire to project an image of

“intergenerational solidarity,” revealing staffers’ sensitivity to accusations that the unions

31

protect pensioners at the expense of everyone else. But this sensitivity is not all hype.

Pensioners’ union officials whom we interviewed were well aware that the Italian pension

system can survive only if it continues to be funded by a new generation of workers, and this

means redirecting social spending towards programs that prepare Italian youth for

employment, stimulate job growth, and make it easier for working parents to raise children.

These concerns appear and reappear in the pages of the journals and newsletters sponsored by

these organizations, which are widely read by professionals who work in the area of social

policy in Italy. The existence of independent pensioners’ unions in Italy has facilitated the

development of a professionalized cadre of union officials who take the long view when it

comes to pension reform.

Finally, sporadic concertation in Italy reveals that relatively high levels of confederal

authority must be available in order for independent pensioners’ unions to facilitate reformist

positions on the part of the confederal leadership. When the confederal unions’ leadership

have been incorporated into the process of reform, reforms tend to go more smoothly and

both category unions and political parties with close ties to unions are less tempted to reach

out to their militant base to gain consensus. The reform phases of 1968-71 and 1994-1996,

which are chronicled in detail by Regini and Regonini (1981) and Baccaro (1999 and 2000)

clearly illustrate this dynamic.

It should be noted, of course, that our characterization of the Italian case has relied on

a generalization about the behavior of confederal and category unions across the three main

confederations. In truth, the CGIL, the largest of the three confederations, is the most

faithfully described by this portrait. The left-leaning CGIL, as a confederation, has tended to

see itself as a “class union” to a greater degree than either the Catholic CISL or the

independent UIL, both of which are closer to AFL-CIO-type “craft” unions. This means that

the CGIL’s confederal leadership’s “natural” inclination is to represent encompassing claims

32

where it can, whereas the CISL and UIL leadership, even at the confederal level, are more

comfortable championing the particular interests of category unions. Still the pensioners’

unions of all three confederations have worked together, developing a common,

encompassing vision of what to do about the welfare state. Particularly during periods of

intense concertation, the pensioners’ unions of the CISL and UIL have been able to push this

vision strongly even on confederal leaderships that might be less likely than the CGIL’s to

adopt an encompassing vision of reform.

During the most recent episodes of reform, Italian pensioners' unions have not staked

out a radically different position from the confederations. This makes it difficult, in the

absence of an ethnographic investigation into the process of union decision-making, to

measure the precise degree of “influence” (see Hacker and Pierson 2001) that pensioners’

unions have over confederal positions. One thing that is certain is that the shifting

environment in which these reform debates have taken place of the last several years is likely

to affect the ability of pensioners’ unions to translate their demands into the confederations’

demands. The difficulties of negotiating reform with the right-wing Berlusconi government

have brought cross-confederal cooperation among the CGIL, CISL and UIL under strain, and

the CISL in particular has reverted to its “natural” tactic of pressing for a particularistic

defense of the pension system status quo. The Berlusconi government has also abandoned

the tripartite context that had over the previous eight years so effectively reduced the power

of the category unions and increased the importance of horizontal groupings like the

confederal leadership and the pensioners' unions. As a result, pensioners’ unions are likely in

the near future to find themselves less able to influence the terms of the debate at the

confederal level.

33

B. Germany

Organization of Pensioners within Unions.

German unions are organized into several federations, but the German Trade Union

Federation, DGB (Deutscher Gewerkschaftsbund), with 80-90 per cent of all union members

is by far the dominant federation.16 Today, the DGB has 8 affiliate unions, down from 14 a

few years ago. Unions are organized along industrial lines, and pensioners represent close to

20 per cent of all union members. Despite their large numbers, retired union members have

few opportunities to exert influence commensurate with their numbers. There are no

independent pensioners’ unions; instead, retirees may remain members of their sectoral

union, but they have relatively little direct influence on sectoral union social policy positions

and no direct representation at the confederal level.

German unions have recently gone through a fundamental process of restructuring.

Partly in response to declining membership, union mergers have reduced the number of union

federations and sectoral unions significantly.17 Formal centralization is weak: wage

bargaining follows a federal structure, with several key sectors setting the standard for the

rest of the economy. Silva (2001) calls collective bargaining in Germany "loose

centralization" because sectoral and regional contracts closely follow the standard set by the

lead sector. Despite what appears to be a high degree of informal centralization in wage

bargaining, the confederal level lacks formal and effective authority over affiliates in wage

bargaining and in other issue areas.

The DGB has the formal power to represent the social policy interests of its affiliates,

but this characterization is deceptive. Although the DGB functions as opinion leader and

16 On German unions and industrial relations, see Thelen 1991; Swenson 1989; Hassel forthcoming; Streeck and Hassel 2003; Ebbinghaus 2002. 17 Until ver.di was created in 2001, there were five union confederations, the DGB (German Trade Union Federation), the DAG (now incorporated into ver.di) represented white collar workers, with 4.5% of all union members; the DBB represents civil servants with 10.6%; the CGB, a Christian union federation, represents 2.8% of all unionized wage earners, and the ULA represents managerial employees, with less than one percent of all union members.

34

mouthpiece for member unions, formulating a common social policy position is not without

difficulties since the DGB's positions are necessarily the product of negotiations among the

member unions. And although the affiliates formally accept the social policy leadership of

the DGB they are free to articulate their own positions. Programmatic differences between

member unions concerning pensions are particularly large given that the unions represent the

interests of clienteles with very different characteristics.

IG Metall (the metalworkers union) has traditionally dominated German industrial

relations and the DGB. Until two years ago, IG Metall was the largest union in the DGB, but

now is surpassed by ver.di (Vereinte Dienstleistungsgewerkschaft), the multi-branch service

union formed in 2001 by the merger of five service sector unions. The remaining six DGB

unions are concentrated in the chemical, building, and printing sectors and are relatively

small compared to IG Metall and ver.di. Of these smaller unions, the IG Chemie and

Bergbau, which is also the product of a merger a few years ago, is arguably the most

influential (see Table 4).

As noted, retirees make up a significant share of union members and the number is

growing. In 1991, 13 per cent of DGB union members were retirees, while today almost 20

per cent of union members are retired (Hassel and Leif 2002). In 1991, there were 14 DGB

unions and the two largest unions, IG Metall and ÖTV (public transport, a forerunner to

ver.di) had 17 per cent and 14 per cent retired members respectively (Künemund 1994b, 33).

The ageing of the membership is especially apparent in IG Metall. In 2003, IG Metall has 2.6

million members, and 570,000 are retired. 1.1 million members (including retirees) are older

than 50 (Frese and Weidenfeld 2003). As Wolf et al. (1994) put it, "measured by the number

of members who are retired, they (unions) are among the largest organizations for retirees."

(12).

35

The relatively high number of retired members has not led to much formal influence

for retirees within the unions (von Winter 1997, 173; Wolf, Kohli and Künemund 1994, 13).

At the union level pensioners' influence ranges from nearly non-existent, to weak

institutionalization. Von Winter (1997,174) argues that the institutions for the representation

of pensioners within unions are not very important, largely because retired unions members

are inactive in union work, pensioners committees in unions are mainly involved in

informational and recreational activities, and pensioners have few opportunities to influence

decision-making. Moreover, institutions for the representation of retirees in the sectoral

unions often operate at the local and regional level, rather than the national level.

A brief examination of pensioner representation in the two largest DGB unions, IG

Metall and ÖTV (now incorporated into ver.di) illustrates the variation in pensioner influence

at the sectoral union level. In the early 1990s, ten of the DGB's 14 unions included

provisions for independent senior organizations in their statutes or organizational guidelines,

including IG Metall and ÖTV (Künemund 1994b, 42). The most common form of

organization for retired union members is the "seniors working group"

(Seniorenarbeitskreis). In 1989 IG Metall adopted a directive (Richtlinie) aimed at

increasing the integration and representation of retired members. The directive requested IG

Metall's administrative units (Verwaltungsstellen)18 to establish a seniors working group if

one did not exist already. However, many working groups were given only advisory status at

the local level, and few financial resources. Moreover, pensioners were not represented on

the national governing board, nor did they have the right to send representatives to the

national congress (see also von Winter 1997, 173). Künemund (1994a) reports that the

activities of the seniors groups depends largely on the role of the retirees active in union

activities, and these tend to be the same people who were active in union work while they

18 In the early 1990s there were 159 administrative units in 7 districts.

36

were working. All of this leads to a wide variation in the level of activity among the seniors

groups.

The weak institutional position of retirees within IG Metall means that they have very

little influence on union politics, including the articulation of social policy preferences. A

large number of retirees remain IG Metall members, but membership seems to be motivated

more by the selective incentives that go along with membership (inexpensive insurance;

access to organized recreational activities, etc) than by the desire to engage in union politics.

Even when retired members express high levels of interest in union politics (add cite), they

have few opportunities for real influence. Moreover, the scarcity of organizational resources

for retirees in IG Metall means that retirees are highly likely to view themselves more as

retired metalworkers with interests largely in line with employed metalworkers than they are

to identify themselves as pensioners with preferences that are related more to retirement than

to sectoral issues. Thus the high number of retirees among IG Metall members remains a

more or less latent group with little influence on IG Metall's core wage earner preferences:

higher wages, generous pension, improved access to early retirement, work time reductions,

etc.

In contrast to IG Metall, pensioners within ÖTV (public sector transport), one of the

forerunners to ver.di, had more institutional opportunities to influence union politics. ÖTV

statutes allowed for the establishment of pensioners committees (Seniorenausschüsse) for

providing advisory and informational services to retired members and to promote the

integration of retirees in the union. In addition, the pensioners committees were represented19

on the local and regional governing committees (Vorstände) as well on the central governing

board (Hauptvorstand)(von Winter 1997, 173). These provisions for the representation of

retirees also exist in ver.di, the successor to ÖTV. Ver.di's 460,000 pensioners (out of a

19 These were unpaid positions (ehrenamtlich).

37

membership of 3.2 million members) are represented at all four organizational levels (local,

regional, district, and national). Retirees are represented at the national congress, but like the

Italian pensioners’ unions, there are self-imposed limits to pensioner representation:

pensioners are content with half of the representation that their numbers would otherwise

give them.

Despite the relatively high level of pensioner organization in ver.di (compared to

other unions), retired members are careful to avoid segregation and ghettoization by focusing

solely on issues related to retirement. The pensioners committees are active in all aspects of

union politics, even if there is more emphasis on retiree-specific issues. The fact that retired

members co-exist with active members probably means that their demands are toned down---

they have less incentive to pursue a pure "pensioner" agenda. Nevertheless, according to one

official, retired members are prepared to accept small pension cuts as long as the burden of

cuts/increased contributions is distributed across age cohorts.20 Here the rationale is that it is

better to accept some cuts in order to make sure that the pension system does not collapse.

Although pensioners have more institutional avenues to influence union policy in

ÖTV/ver.di, this has not necessarily translated into a more pensioner-oriented agenda for the

union. Despite their numbers and representation within the union, pensioners in ÖTV/ver.di

appear to remain oriented toward sector-specific wage earner issues, even if they do make

their voices heard on retirement-specific issues. Thus even when pensioners are organized

within a sectoral union and have access to union decision-making, they do not automatically

pursue a "pensioners" agenda.

The dominance of wage earner interests over pensioner interests is even stronger at

the confederal level, where retired union members have very little institutional influence.

20 Interview with ver.di official, 29 July 2003.

38

For example, the statutes of the DGB provide for a national level committee for women's

issues and youth issues, but nothing similar for pensioners' issues (DGB 2002). The only

avenue for formal influence by retirees is through their own sectoral unions.

In sum, German unions have a significant number of retirees among their members,

institutions for the representation of pensioners within the unions are weak, and the wage

earner interests of sectoral unions dominate union politics. Given that unions are the primary

organizations representing German pensioners, at least in terms of membership numbers, the

dominance of wage earner interests within the unions is surprising. Unions are no doubt

aware that the rising number of retired members has the potential to lead to internal conflict

between employed and retired members, which probably explains why pensioners have so

little access to union decision-making and are not encouraged to trade in their sectoral

(former) wage earner identity for a pensioner identity.

Unions and Pension Policymaking.

The German pension system is structured according to occupational status, benefits

are earnings-related, and civil servants have a separate, tax-financed pension scheme. The

self-employed and professionals are covered by their own pension funds, while the statutory

pension insurance (GRV, Gesetzliche Rentenversicherung) covers blue collar and white

collar workers, about 70 per cent of the workforce. Statutory pensions are financed by social

insurance contributions divided evenly between employers and employees, and financing is

PAYGO. The revenue from contributions is not enough to cover expenditures, so the federal

government kicks provides a subsidy that has grown steadily during the past decade. The

subsidy currently covers one third of statutory pension spending.

Several trends have combined to create substantial financial pressure on the pension

system. First, rising unemployment and early retirement have reduced the contribution base

39

and increased costs. In addition, many of the costs of German unification have been financed

via the social insurance system, including pensions. These trends resulted in continuously

rising contribution rates and predictions of unsustainable contribution levels (40 per cent)

without reform.

Until the 1990s, pension policymaking in Germany was consensual and dominated by

a closed corporatist policy network consisting of the DGB, BDA (employers), the two main

organizations for administering pensions, VDR (Verband Deutscher

Rentenversicherungsträger) and BfA (Bundesversicherungsanstalt für Angestellte), as well as

the pension policy experts in the ministries and political parties. The role of the employers

and unions was further strengthened by the corporatist structure of the VDR and BfA.

According to von Winter (1997), this closed policy network largely decided the main features

of pension policy development until the end of the 1980s.

This pension policy network had the following rough division of labor: the unions

and employers (DGB and BDA) would try to work out a compromise position on pension

issues because of their joint interest in "peace" in this policy area. As the administrators of

the statutory pension scheme, the VDR and BfA possessed expertise on pension issues and

provided a forum for negotiation because of their bipartite structure. This closed community

produced pension compromises for several decades because the policy positions of the social

partners were not widely divergent. So this policy network typically worked out a

compromise on its own and then sold it to the politicians in government and parliament.

During the heyday of the closed pension network, the DGB was usually able to assert

its authority over the affiliate unions and work out a unified union position. For example, the

DGB negotiated on behalf of the unions for the 1992 reform, and it supported some reform

provisions (limits on early retirement) that IG Metall opposed. As long as inter-union

differences in policy preferences were not large, this arrangement functioned reasonably well.

40

However, pension financing became a recurring problem during the 1990s, driving the

programmatic differences among unions and between unions and employers further apart.

This made it harder for the DGB to push a unified position, and it made it harder for the

social partners to agree.

Union structure and confederal level positions concerning the 2001 pension reform

By the mid 1990s, financing problems pushed the pension reform onto the political

agenda again. However, this time the cross-party consensus had evaporated and the

corporatist policy community did not play as strong a role as in the past. In 1997, the

Christian Democratic-Liberal coalition passed an unpopular pension reform that introduced

modest cuts for future retirees. Initially the government attempted to secure opposition and

union support, but these efforts failed. Although the proposed cuts were slated to affect

future pensioners, the DGB and sectoral unions opposed them, arguing instead that reform

should focus on improving the financing of pensions. Unmoved by these protests, the

government went ahead with the reform. The long consensus on pensions was now broken.

The 1998 election ended more than a decade of CDU-FDP rule and brought a Social

Democratic-Green Party coalition to power. Pensions were a major election issue and the

SPD reversed some of the cuts adopted by the Kohl government in 1997. Expectations ran

high that the government would adopt a more moderate pension reform than the 1997 reform,

but the government's first proposals dampened these hopes. In early 1999, the government

announced its reform plans, including more drastic cuts than the ones the government had

just reversed. The initial proposal included cuts in future benefits and obligatory private

provision to fill the gap caused by reduced benefits. Current pensioners and those retiring

before 2011 would be spared.21

21 The details of the 2001 reform are based on coverage in the Süddeutsche Zeitung.

41

The red-green coalition tried to get union support for their plans but quickly

encountered strong opposition. IG Metall opposed the reform, ÖTV (now ver.di) expressed

sympathy with the goals of the reform but opposed some elements of it, and the DGB had

difficulty voicing a unified position. The conflict continued for several months, with union

demonstrations, etc., until the chancellor held talks with the sectoral and confederal union

leaders to hammer out a compromise. The government made some changes in its proposal,

and parliament passed a watered down version in January 2001.22

The government's original proposal concentrated cuts on future pensioners (retiring

after 2010) via the "Ausgleichsfaktor" or "compensation factor." The idea was to protect the

pensions of current pensioners and those retiring before 2010. In 2011, the new formula

would kick in, and post-2011 pensioners would receive lower pensions.23 The authors of the

reform calculated that wage earners close to retirement and current retirees would not have

time to accumulate alternative means of supplementing their pensions, while those retiring

after 2011 would have ample time to make changes in their retirement planning (cf. Ebert

2002). Funded occupational and private pensions (encouraged by tax breaks) would replace

the income lost by the pension cuts for those retiring in 2011 and afterwards.

The original reform formula seems to have been tailor-made to appeal to the

organized elderly, if such a group were available to lend its support. The proposal spared

current pensioners, thus limiting potential electoral retribution by elderly voters, and it

concentrated cuts on future pensioners, who as current wage earners would be offered

incentives to make up for public pension cuts by increasing their private pension savings.

The proposal's attractiveness to the elderly seems to have mattered little, however, probably

because unions are the primary representatives of pensioners' interests in Germany, and 22 The bill was split into two parts to faciliate parliamentary passage. The Bundestag adopted the legislation in January 2001 while the Bundesrat adopted a separate part of the reform in May 2001. See Anderson and Meyer (2003) and Schludi (2001) for details. 23 Between 2011 and 2030 pensions would decrease by 0.3% annually for a total decrease of 6%. The final pension level would be 64% of average net wages compared to the current pension level of 70% of net wages.

42

pensioners have little independent influence in the unions. Instead, a coalition consisting of

the left wing of the SPD, the pension organizations (VDR and BfA), and the unions

succeeded in getting the government to change the proposal. The VDR and BfA pushed for a

generation-neutral reform that would spread smaller cuts across more age cohorts (including

pensioners)24; the SPD Leftists pushed for provisions to limit the cuts for younger workers;

and the unions pressed for a reform with the smallest possible cuts. This coalition proposed

less drastic benefit cuts (67% of net wages instead of 63%), slightly higher contribution rates

than those envisaged in the first proposal, and incentives for occupational pensions. The

government's final reform was based on these demands.

How did the organization of pensioners influence the DGB's position and union

bargaining in the reform process? First, the original reform proposal largely spared current

pensioners and anyone retiring before 2010, while the final compromise spread the burden of

smaller cuts across current workers and all pensioners beginning in 2011. Pensioners will

take a small hit (starting in 2011), but they still feel more pain than in the original proposal.

The unions were a major force behind this change. This outcome is surprising given the

conventional wisdom about the electoral clout of elderly voters and the supposed effects of

the "seniority bias" thesis. Clearly, the final compromise reflects wage earner interests rather

than pensioner interests. If there had been an independent pensioners union like in Italy, we

might have seen this union attempt to push the DGB toward a more substantial (in terms of

achieving financial balance) reform. Instead, a central concern of the unions (and their allies

in the SPD and pension administration) was to protect the pensions of future retirees, i.e.

current workers. Above all, the unions wanted to make sure that future pensions would

continue to deliver benefits in line with the standard of living achieving during employment.

This principle is deeply entrenched in union preferences concerning social policy, and the

24 In the final compromise, all pensions would decrease beginning in 2010, including the pensions of those who retired before 2010.

43

magnitude of the proposed cuts threatened to undermine it. Unions wanted to prevent this at

all costs, which is why they were willing to spread the burden of cuts by applying them not

just to those retiring after 2010, but to all pensioners. That many union members are retired

and would have been better off with the original proposal made little difference; retired union

members have little organizational clout and their preferences as retired union members

remain unarticulated.

Second, there were significant differences among the sectoral unions about the

reform: IG Metall was largely opposed to the reform and preferred instead to raise

contributions or to broaden the revenue base. In the midst of the pension reform debate, IG

Metall even proposed a separate scheme for retirement at 60 that ultimately failed. ÖTV, the

public sector and transport union, was one of the most accommodationist unions. ÖTV has

one of the most institutionalized forms of representation for retired members, but additional

research is required before we know if this mattered for the ÖTV's reform stance.

Differences among the sectoral unions made it difficult for the DGB to adopt a unified

position, to moderate category-specific pension interests, or to speak with authority for its

affiliates. Indeed, the reform debate was often chaotic, with sectoral unions even engaging in

open criticism of each other's positions. The DGB's relative lack of authority over its

affiliates also contributed to the dominance of sectoral pension interests.

C. Italy and Germany compared

To what extent do the Italian and German cases support the theoretical propositions

advanced in this paper? Our central hypothesis is that union confederations are likely to

support pension reform proposals that concentrate cuts on future beneficiaries rather than

current pensioners when two conditions are present: there is a high potential for pensioner

influence at the confederal level, and the confederal level has authority over sectoral unions.

44

Our two cases vary on both dimensions. In Italy, pensioners can have an influential voice at

the confederal level because they are organized in independent pensioners’ unions.

Additionally, under conditions of concertation, Italian union confederations have the capacity

to speak authoritatively for their affiliates. In sum, in Italy the conditions that we hypothesize

to facilitate pension reform cooperation by peak union organizations are present under certain

conditions. In contrast, German union structure means that retired members' interests as

pensioners remain unarticulated within unions because pensioners have very little influence

on union social policy-making. Instead, the interests of wage earners in stable employment

prevail within the sectoral unions. Moreover, confederal authority over affiliate unions is

weak, so uncoordinated bargaining is likely. In sum, Italy scores high on both independent

variables while Germany scores low.

What happened when pension reforms actually reached the political agenda and union

leaders faced the decision of whether to support or oppose reforms? Are the implications of

our theory observable in the cases studies?

Our theory predicts that Italian pensioners’ unions should articulate interests based on

their status as pensioners and not on their former status of wage earners in a particular sector.

Pensioners and wage earners have different preference orderings concerning pension reform,

so we should observe a broad pattern of pensioners’ unions advancing pensioner interests that

are distinct from wage earner interests. In keeping with their interest in the sustainability of

the pension system and their desire to keep benefits at least at current levels, pensioners’

unions should support reforms designed to improve financial sustainability that concentrate

losses on future rather than current pensioners. This broad pattern of divergent interests is

present in all of the reform episodes examined in this paper, including the ones that failed. In

the reform discussions of 1968, 1976, 1978, 1992, 1995, and 1996 pensioners’ unions

supported reforms that cut future benefits, raised contributions for current workers, and/or

45

increased the state subsidy to the pension. Not all of these reforms were successful; indeed

opposition by strong sectoral unions derailed the reforms proposed in 1968 and 1978. But the

pensioners’ unions consistently promoted reform strategies designed to improve pension

system sustainability by concentrating losses on current workers, either through higher

contributions or lower future pensions, or both. To be sure, pensioners’ unions had little to

fear in terms of reduced benefits. But given that it would have cost the pensioners’ unions

little to support the opposition of sectoral unions in the reform discussions that failed in 1968

and 1978, the pensioners’ unions' more moderate stance is striking.

A second implication of our theory is that the independent voice of pensioners at the

confederal level should matter for the positions that confederal unions take on pension

reform. Our review of the Italian evidence also provides support for this implication of the

theory. Pensioners’ unions help to sway confederal leaders in favor of pension reforms in all

of the reform episodes reviewed here. However, the confederal leadership was not always

able to overcome opposition on the part of sectoral unions. This leads us to the third

implication of the theory: when the confederal level is able to control its affiliates and speak

with one voice, a pro-reform stance is more likely. In the Italian reform episodes in which

concertation occurred, union confederations' authority over affiliates increased, and the

confederal leadership was able to override the opposition of sectoral unions to proposed

pension reforms. And as discussed, the influence of pensioners’ unions was crucial to this

process.

In sum, the reform episodes of 1968, 1978, 1992, 1995, and 1996 provide broad

support for the theoretical arguments advanced in this paper. The reform proposals differed

in each of these episodes, but the general pattern is clear: pensioners’ unions consistently

supported reforms that did not cut current pensions, and their support helped confederal

leaders to overcome opposition from strong sectoral unions in reform negotiations.

46

Does our theory also hold in the German case? As argued, German union structure

priveleges wage earner interests over pensioner interests, and confederal authority over

affiliates is weak, increasing the likelihood of uncoordinated bargaining.

First, we hypothesized that the organizational representation of pensioners in German

unions leads to the privileging of wage earner interests over pensioner interests, despite the

large number of retired union members. We argued that this should be especially true in

strong sectoral unions, and this was indeed the case concerning IG Metall's role in the reform

process. Retired members of IG Metall have little influence on union policies, so the IG

Metall has every incentive to defend the wage earner interests of its members. By taking an

oppositional stand in the reform negotiations, this is precisely what IG Metall did

Second, we argued that strong sectoral unions in which pensioners have

organizational influence are likely to advocate more moderate positions, even though these

positions still largely reflect wage earner interests. Here, the institutionalized influence of

retired members, pursuing their interests as pensioners and former wage earners, should

moderate sectoral union demands. The policy preferences and behavior of ÖTV in the 2001

reform appear to support this hypothesis, but additional research is required in order to

ascertain whether retired members in ÖTV were a moderating force, or whether some other

factors played a role.

Finally, we argued that the absence of an independent pensioners' voice at the

confederal level decreases the likelihood that confederal leaders can overcome opposition by

strong sectoral unions to proposed pension reforms that impose losses on current workers.

During the 2001 reform process, the German DGB broadly supported the goal of pension

reform, but opposition to the government's proposals by strong sectoral unions (especially IG

Metall) constrained the DGB's capacity to advance a unified position. The most striking

47

aspect of the reform process from our point of view is that wage earner interests dominated,

despite the presence of large numbers of retirees among union members.

IV. Conclusion

The central task of this paper has been to elaborate the conditions under which

confederal leaderships will adopt positions favorable to pension system retrenchment, even

when there are large numbers of pensioners within the unions. Whereas most scholars have

assumed that ageing union memberships will inexorably lead labor unions to the intransigent

defense of the pension system status quo, we argue that large numbers of retired members do

not necessarily lead confederal union leaders to oppose pension system reforms. In fact, we

expect that confederal leaders are likely to agree to pension system reforms under the

following three conditions: 1. pensioners have a strong institutional presence at the

confederal level; 2. the confederal level has a high degree of authority over affiliates; and 3.

proposed reforms affect future benefits rather than current pensions.

This expectation is based on a number of important priors, which may be stated as

assumptions that may or may not be perfectly accurate with regard to any individual county

case.

• Following Olson (1982), more encompassing union structures (the confederal

leadership, unions representing workers or pensioners drawn from multiple sectors)

will be more likely to support pension system retrenchment than sectoral unions –

particularly strong sectoral unions whose members enjoy extensive pension rights.

• The confederal leadership is more likely than category unions to support pension

system retrenchment during periods of concertation, because the confederal leadership

wants to maintain its position as a valued interlocutor for government and employers.

(This assumption might easily be altered, though, depending on the partisan

48

composition of government, with unions being less interested in cooperation when the

Right is in government.)

• Individual pensioners care more about cuts to their own benefits than cuts to future

benefits. At the same time, the current generation of pensioners is likely to have less

at stake in the level of pensions per se (as opposed to social services, health care,

etc.), since a substantial number of current pensioners matured their pension rights at

a time when benefits were less generous than they are for workers retiring today. The

implication of these assumptions is that current pensioners (and their representatives)

are likely to be less concerned than active workers about pension system

retrenchment, especially if the level of social provision to current pensioners is

maintained.

We do not seek in this paper to substantiate empirically these prior assumptions. But

analysis of the behavior of unions during episodes of pension reform negotiations in Italy and

Germany does support our main hypotheses about the conditions under which unions will be

likely to support pension system reform. In Italy, strong, independent pensioners’ unions

aggregated the diverse preferences of multiple generations of pensioners into a moderate pro-

reform position. Representatives of the pensioners’ unions used their expertise on welfare

state issues and their privileged relationship with confederal leaders to influence the

confederations’ positions on pension reform. In Germany, by contrast, pensioners remained

organized within the category unions representing the sector of their former employment, and

as a result could contribute little to confederal position on pensions. Union positions on

pension reform remained dominated by the concerns of active workers in strong category

unions. In the German case, it seems likely that independent pensioners’ unions could have

49

strengthened the confederal leadership’s efforts to reform the pension system through

revisions of the pension rights of future beneficiaries, rather than cuts to current pensioners.

Our argument challenges much of the conventional wisdom about elderly interest

groups as well as the theoretical arguments about interest groups and unions with ageing

memberships. Unions containing large and well-organized pensioner “lobbies” are not, we

argue, likely to be the most important “obstacles” to pension system reform in Europe. In

fact, evidence from Italy and Germany suggests that independent pensioners’ unions and

sectoral unions with independent pensioner influence are more likely to support reforms than

are strong sectoral unions dominated by current workers. The “graying” of union

membership is then unlikely to be the source of union intransigence on pension reform. This

implies that strengthening the representation of pensioner interests within unions may

actually help, rather than hinder, the process of reform.

The arguments and analysis presented in this paper are a first attempt to address a

larger theoretical problem: where do union preferences regarding pension reform come from?

We think that there are compelling reasons to believe that the organization of pensioners

within unions makes a difference in unions' attitudes towards pension reform. Clearly, other

factors matter as well. Disentangling the effects of the number of pensioners, pensioner

organization, union centralization, the intensity of tripartite concertation, union control over

pension system resources, macroeconomic conditions, the presence of non-union groups

representing elderly interests, etc. is our task as we move forward with this research.

50

Bibliography Acoccella, Pino. 1999. Un sindacato degli anziani capace di parlare ai giovani. Oggidomani

Anziani 12 (2). Anderson, Karen M. 2001. The politics of retrenchment in a Social Democratic welfare state.

Reform of Swedish pensions and unemployment insurance. Comparative Political Studies: 34, 9, 1063-1091.

Anderson, Karen M. and Traute Meyer. 2003. Social Democracy, Unions, and Pension

Politics in Germany and Sweden, Journal of Public Policy: 23 (2), 23-54. Baccaro, Lucio. 1999. The Organizational Consequences of Democracy: Labor Unions and

Economic Reforms in Contemporary Italy. PhD diss., Sloan School of Management and Department of Political Science, Massachusetts Institute of Technology, Cambridge, MA.

-----. 2000. The Italian Experience in European Perspective. Paper prepared for the 12th

International Conference of Europeanists, Chicago, March 30-April 2, 2000. Baldissera, Alberto. 1996a. Conservare o cambiare il sistema pensionistico italiano? ISPO

Political Trend 12. -----. 1996b. La rivolta dei capelli grigi. Il caso italiano e francese. In Il paese dei paradossi.

Le basi sociali della politica in Italia, edited by N. Negri and L. Sciolla. Rome: Nuova Italia Scientifica.

Brugiavini, Agar, Bernhard Ebbinghaus, Richard Freeman, Pietro Garibaldi, Bertil Holmund,

Martin Schludi, and Thierry Verdier. 2001. Unions and pensions: theory, evidence and implications. In The Role of Unions in the Twenty-First Century: A Report for the Fondazione Rodolfo Debenedetti, edited by T. Boeri, A. Brugiavini and L. Calmfors. Oxford: Oxford University Press.

Campbell, Andrea, and Julia Lynch. 2000. Whose 'Gray Power?' Elderly Voters, Elderly

Organizations, and the Welfare State in Italy and the United States. Italian Politics and Society 53 (Summer).

Cazzola, Giuliano. 1998. Pensioni d'anzianita': Iniquo peso contro i giovani. Il Sole 24 Ore,

September 3, 1998, 1. DGB. 2002. Satzungen. www.dgb.de. Dickson, Eric and Kenneth Shepsle. 2001. Working and Shirking: Equilibrium in Public

Goods Games with Overlapping Generations of Players. Journal of Law, Economics & Organization 17: 285-318.

Ebbinghaus, Bernhard. 2002. Trade unions’ changing role: membership erosion,

organisational reform, and social partnership in Europe. EU Paper Series, The European Union Center, University of Wisconsin-Madison.

51

Ebbinghaus, Bernhard and Jelle Visser. 2000. The Societies of Europe. Trade Unions in Western Europe since 1945. London: Macmillan. Ebert, Thomas. 2001. Rentenreform 2001: Sozialverträgliche Modernisierung? Sozialer

Fortschritt 8/2001: 182-187 Ferrera, Maurizio. 1984. Il welfare state in Italia: Sviluppo e crisi in prospettiva comparata.

Bologna: Il Mulino. -----. 1996. Cronache del welfare: Queste pensioni sono la minaccia dei giovani. Il Sole 24

Ore, June 28, 1996, 7.

FNP-CISL. 1992. Corsi di formazionae di base -- Breve storia della FNP. Vol. 2. Rome: FNP-Cisl.

Frese, Anton and Ursula Weidenfeld. 2003. Die Prinzen Eisenherz. Tagesspiegel July xx 2003.

Hacker, Jacob, and Paul Pierson. 2001. Business Power and Social Policy: Employers and the

Formation of the American Welfare State. Unpublished paper, Cambridge, MA: Harvard University.

Hall, Peter. A. and Soskice, David. 2001. Varieties of Capitalism - The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press.

Hassel, Anke. forthcoming. Organisation: Struktur und Entwicklung. In: Wolfgang Schröder

and Bernhard Weßels (eds.): Gewerkschaften in Politik und Gesellschaft der Bundesrepublik - Ein Arbeitsbuch. Opladen: Westdeutscher Verlag.

Hassel, Anke and Thomas Leif. 2002. Reformfähige Gewerkschaften - der Zukunftsprozess

der IG Metall. Frankfurter Rundschau. May 29, 2002. Institut des sciences du travail, Université catholique de Louvain. Social Organizations in

Europe : European Trade Union Confederation - ETUC/CES. http://www.trav.ucl.ac.be/partenaires/eu-8-en.htm.

Künemund, Harald. 1994a. Ausmaß, Verbreitung, und Organisationsformen-Das Spektrum

der IG-Metall-Seniorenarbeit, in Jürgen Wolf, Martin Kohli, and Harald Künemund, eds. Alter und gewerkschaftliche Politik.

-----. 1994b. "Senioren im Deutschen Gewrkschafstbund," in Jürgen Wolf, Martin

Kohli, and Harald Künemund, eds. Alter und gewerkschaftliche Politik. Künemund, Harald, Sighard Neckel, and Jürgen Wolf,. 1994. Der Raum der Möglichkeiten -

Organisationsalternativen der gewrkschaftlichen Seniorenarbeit. in Jürgen Wolf, Martin Kohli, and Harald Künemund, eds. Alter und gewerkschaftliche Politik.

Levy, Jonah. 1999. Vice into Virtue? Progressive Politics and Welfare Reform in Continental

Europe. Politics and Society 27:2, 239-273.

52

Lynch, Julia. 2002. The Age of Welfare: Citizens, Clients, and Generations in the Development of the Modern Welfare State. PhD diss., Dept. of Political Science, University of California, Berkeley.

Maestri, Giovanni Ezio. 1994. Rappresentanza degli interessi, partiti e consenso: Giungla

pensionistica, clientelismo e competizione politica in Italia. Milano: FrancoAngeli. Manghi, Bruno. 1994. Il sindacato deve tener conto di tre generazioni con esigenze

contrastanti; Quando l'eta' e' anche simbolo. Il Sole 24 Ore, September 30, 1994, 7. Manow, Philip. 2001. Comparative Institutional Advantages of Welfare State Regimes and

New Coalitions in Welfare State Reforms. In P.Pierson (ed.)., The New Politics of the Welfare State. Oxford: Oxford University Press.

Mares, Isabela. 2003. The Sources of Business Interest in Social Insurance: Sectoral versus

National Differences. World Politics. Massarenti, Armando. 1997. Il welfare delle categorie privilegiate danneggia le prospettive

delle generazioni future; Troppo garantiti contro poco protetti. Il Sole 24 Ore, December 6, 1997, 4.

Olson, Mancur. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and

Social Rigidities. New Haven: Yale University Press. Pampel, Fred C. 1994. Population Aging, Class Context, and Age Inequality in Public

Spending. American Journal of Sociology:153-195. Pedersini, Roberto. 2000. Pensioners' trade union organisations examined.

http://www.eiro.eurofound.ie/2000/09/feature/IT0009272F.html. Pontusson, Jonas. 1992a. Introduction: Organizational and Political-Economic Perspectives

on Union Politics. In M. Golden and J. Pontusson, eds... Bargaining for Change. Union Politics in North American and Europe. Ithaca: Cornell University Press.

-----. 1992b. The Limits of Social Democracy. Investment Politics in Sweden. Ithaca: Cornell

University Press. Regini, Marino, and Gloria Regonini. 1981. La politica delle pensioni in Italia: Il ruolo del

movimento sindacale. Giornale di diritto del lavoro e di relazioni industriali 10:217-242.

Regonini, Gloria. 1987. Le politiche pensionistiche. In Le relazioni fra amministrazione e

sindacati, edited by I. p. l. S. D. A. Pubblica. Milano: Giuffre. -----. 1990. La politica delle pensioni. In Le politiche pubbliche in Italia, edited by B. Dente.

Bologna: Il Mulino. -----. 1996. Partiti e pensioni: Legami mancanti. In Il gigante dai piedi di argilla: La crisi del

regime partitocratico in Italia, edited by M. Cotta and P. Isernia. Bologna: Il Mulino. Rossi, Nicola. 1997. Sindacati troppo cauti sui privilegi. Il Sole 24 Ore, August 23, 1997, 3.

53

Schludi, Martin. 2002. Pension reform in social insurance countries. MPI Discussion Paper. Silva, Stephen J. 2001. Every Which Way but Loose. German Industrial Relations since

1980, in Andrew Martin and George Ross, eds., The Brave New World of European Labor. European Trade Unions at the Millennium. New York: Berghahn Books.

SPI-CGIL. 1996. Lo Spi-Cgil: I suoi iscritti, le sue strutture. Rome: Editrice LiberEta'. Swenson, Peter. 1989. Fair Shares. Unions, Pay and Politics in Sweden and West Germany.

Ithaca: Cornell University Press. -----. 2002. Capitalists Against Markets. Oxford: Oxford University Press. Thelen, Kathleen. 1991. Union of Parts. Labor Politics in Postwar Germany. Ithaca: Cornell

University Press. Visser, Jelle. 1990. In Search of Inclusive Unionism. Bulletin of Comparative Labour

Relations 18. Winter, Thomas von. 1997. Sozialpolitische Interessen. Konstituierung, politische

Repräsentation und Beteiligung an Entscheidungsprozessen. Baden-Baden: Nomos.

54

Table 1: Pensioners as a percentage of all union members, ca. 2000 Country Confederation Pensioners

(% of all union members) Source

Austria ÖGB 17 IST ÖGB 19 Adam p.c. Belgium FGTB 7 IST CSC 10 IST Denmark LO 3 IST Finland SAK 15 Ahtiainen p.c. STTK 8

Ahtiainen p.c.

AKAVA 8 Ahtiainen p.c. France CFDT 13 IST CFDT 28 Labbé p.c. CGT 22 IST CGT 32 Labbé p.c. FO 59 IST CFTC 6 IST Germany DGB 19 Streeck, 2001 Greece No data Ireland ICTU 2 IST Italy CGIL 55 Pedersini CISL 50 Pedersini UIL 25 Pedersini Luxembourg CGT-L No data LCGB 14 IST Netherlands FNV

9 Centraal Bureau voor de Statistiek, NL

CNV 13 Centraal Bureau voor de Statistiek, NL MHP 9 Centraal Bureau voor de Statistiek, NL Norway LO 22 Nergaard p.c. Poland Total 10 Zadura p.c. Portugal CGTP 3 IST UGT 4 IST Slovenia Total 5 Skledar p.c. Spain CCOO 6 IST UGT 6 IST ELA-STV 11 IST Sweden LO 13 Kjellberg p.c. TCO 11 Kjellberg p.c. SACO 6 Kjellberg p.c. UK TUC 5 IST Data from union publications and personal communications with union officials and national industrial relations experts, especially those affiliated with institutions linked to the European Industrial Relations Observatory network. For a list of these organizations, see http://www.eiro.eurofound.eu.int/contact.html. We also relied on a report on the European Trade Union Confederation, prepared by the Catholic University of Leuven’s Institut des Sciences du Travail, which reports membership figures for the pensioners’ unions affiliated with the ETUC’s pensioners’ federation (http://www.trav.ucl.ac.be/partenaires/eu-8-en.html). Figures drawn from the IST report seem approximate at best, though, and we used them only when no other information was available.

55

Table 2: Organization of Pensioners within European Trade Union Confederations Country Pensioners have own

unions? Pensioners have official current at confederal level?

France Yes Yes Italy Yes Yes Portugal Yes Yes Spain Yes Yes European Level

Yes No

Ireland No Yes Luxembourg No Yes Netherlands No Yes UK No Yes? Belgium No ? Austria No No Finland No No Germany No No Norway No No Sweden No No Denmark No No Greece No data

56

Table 3: Visser’s (1990) Confederal Authority Measure Country Confederal authority

vis-à-vis category unionsUK LOW .5 France LOW 1.0Switzerland LOW 2.5Italy MED 3.5Denmark MED 4.0Germany MED 5.0Netherlands MED-HIGH 6.5Sweden HIGH 7.5Norway HIGH 8.0Austria HIGH 9.5

Table 4: Membership in DGB unions Membership statistics (Dec. 31, 2002)

Deutscher Gewerkschaftsbund

Union manual workers

salaried employees

tenured civil servants

youth male female total percent of DGB total

IG Bauen-Agrar-Umwelt (construction)

439,046 48,844 1,912 51,637 423,434 66,368 489,802 6.4

IG Bergbau, Chemie, Energie (mining, chemicals)

654,814 178,879 0 60,988 673,609 160,084 833,693 10.8

Gew. Erziehung und Wissenschaft* (education and science)

0 95,570 122,637 40,357 84,576 180,108 264,684 3.4

IG Metall (metalworkers)

2,187,829 456,144 0 218,256 2,150,527 493,446 2,643,973 34.3

Gew. Nahrung- Genuss- Gaststätten

189,344 56,006 0 23,196 147,524 97,826 245,350 3.2

Gew. der Polizei* (police)

8,022 20,171 156,714 34,849 152,285 32,622 184,907 2.4

TRANSNET* 70,539 39,450 38,132 15,481 236,427 60,944 297,371 3.9ver.di (services) 853,062 1,553,729 229,495 142,303 1,384,235 1,355,888 2,740,123 35.6 DGB total 4,402,656 2,448,793 548,890 587,067 5,252,617 2,447,286 7,699,903 100

* Some unions' statistics include "other" members (self-employed, unemployed, pensioners, etc). These are included in the totals for some unions which is why the individual numbers do not always add up to the totals given for the DGB at the bottom.